A Knowledge-Light Mechanism for Explanation in Case-Based Reasoning
Year: 2005
Type: other
Abstract: Decision support systems are currently achieving higher classification accuracies by using more\ncomplex reasoning mechanisms. Examples of such mechanisms include support vector machines\nand neural networks. However in spite of these increases in accuracy many decision support\nsystems are not accepted by users. In domains where there is a high cost associated with incorrect\nclassifications, such as medical domains, users are not always willing to accept a decision support\nsystem's classification without proper justification.\nIn every walk of life, from the home to the workplace, people use explanations all the time\nto justify their opinions. Explanations can have many different forms depending on the context\nin which they are used. Over the last few decades there has been a vast amount of research by\nphilosophers into the importance and the requirements of suitable explanations.\nIn spite of the importance of suitable explanations to justify an opinion, many decision support\nsystems fall sort of this requirement. Part of the reason for this is that in many types of decision\nsupport systems it is often extremely difficult, if not impossible, to produce explanations. This\nis particularly the case for black box systems such as support vector machines. However in other\nsystems such as rule-based systems, where the explanation can be in the form of a rule, the\nexplanations can often be complicated and result in confusion for users.\nAlternatively Case-based Reasoning (CBR) Systems lend themselves naturally to producing\nexplanations. As the reasoning in CBR systems is performed on the most similar past case(s) to\na current problem, these similar cases can be used as an explanation for a classification. As these\nsimilar cases are real past problems, they are generally easily understood by users.\nIt is our believe however, that in CBR systems, that there are more suitable cases to use as\nan explanation than simply using the most similar cases. It is our belief that these more suitable\ncases lie between the problem case and the perceived decision boundary. This results in the cases\nforming an a fortiori argument. We describe a framework that we have developed for selecting\nsuch cases.\nWe also believe that it is often not enough, regardless of the suitability, to just use a case as\nan explanation. In our framework we included a mechanism for generating explanatory text that\ncan express why the case is suitable, or in some situations aspects of the case that may not be\nsuitable. This explanatory text can further assist users to decide if they agree with the opinion of
Author: Dónal Doyle
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2278610641
